Business

How AI Could Reshape the U.S. Economy and What the Fed Is Watching

AI and the economy
AI and the economy

Federal Reserve Chair Kevin Warsh says Fed staff have started modeling how artificial intelligence could change U.S. growth and employment. In one scenario, AI helps the economy grow faster while employment stays strong. In another, output rises even as unemployment climbs. A third considers what happens if the returns on AI investment disappoint and businesses pull back.

For the Fed, the timing matters. Companies can spend heavily on AI infrastructure today, while the productivity gains that make goods and services cheaper may take years to spread. That gap could influence inflation, hiring, and how policymakers think about interest rates.

What Did Kevin Warsh Say About AI and the Economy?

In Warsh’s opening remarks at the Fed’s Productivity and Jobs Task Force conference in San Francisco on November 16, 2026, he identified a possibility that goes beyond automating routine work. AI could speed up the process of generating and testing new ideas. If that leads to more inventions and better ways of producing goods and services, the economy could sustain a faster pace of growth.

Fed staff have begun using economic models to explore different outcomes. Warsh presented three illustrative scenarios for the next several years and gave no probabilities for them.

ScenarioWhat it could mean for the U.S. economy
Productivity gains spread widelyBusinesses produce more, workers share in the gains, and economic growth runs above its long-term average while employment remains strong.
Output rises as some jobs disappearAutomation boosts production, while displaced workers take time to find roles that match their skills. Growth and unemployment can rise together.
AI investment disappointsReturns fall short of expectations, stock valuations decline, and businesses cut back on investment. Growth slows and unemployment rises.

Why a Faster Task May Take Time to Show Up in GDP

An AI tool might help a customer service team resolve more requests per hour. To see a gain across the whole company, you also have to account for the cost of deploying the tool, checking its answers, and training staff to use it. The same calculation plays out across thousands of businesses, each adopting the technology at a different pace.

Fed Vice Chair Philip Jefferson said on November 13 that U.S. labor productivity growth averaged 2.2% between 2020 and 2025, up from 1.5% over the previous business cycle. He pointed out that the improvement began before popular AI chatbots appeared in 2022. The figures show that productivity picked up, while AI’s contribution to that increase remains uncertain.

Could AI Raise Inflation Before It Lowers Costs?

Jefferson described two forces that operate on different timelines. In the near term, building and equipping data centers increases demand for construction, computing hardware, and other resources. When suppliers struggle to keep up, prices can rise. Jefferson cited memory and logic chips as examples.

That’s one example of how AI demand can affect smartphone prices. Some of the same supply pressures that raise costs for AI infrastructure can reach consumer electronics. Over time, manufacturers can expand supply, and businesses that use AI effectively may be able to produce more at a lower cost per unit. Whether customers see lower prices depends in part on how much of those savings businesses pass on.

What Could AI Mean for Jobs and Interest Rates?

The Fed has to consider why unemployment is rising if that happens. When households and businesses cut spending, lower borrowing costs can help support demand. When automation changes the skills employers need, displaced workers may still need time to retrain or find a new role. The two problems can occur at the same time, and they respond differently to monetary policy.

There’s also a financial risk. Fed Governor Michael Barr warned that if AI investments take longer to pay off than investors expect, falling stock valuations could weaken spending. Debt used to finance data centers could amplify that stress. A technology can improve productivity over the long run while the companies financing its rollout face losses along the way.

Warsh’s remarks gave no timetable for a rate change. What will matter for future decisions is whether productivity gains spread beyond a few industries, whether hiring keeps pace with job displacement, and how quickly supply catches up with AI-related spending.

Filed under: Business

← All articles